Liu et al. Radiation Oncology (2015) 10:185 DOI 10.1186/s13014-015-0493-6
RESEARCH
Open Access
Time-window of early detection of response to concurrent chemoradiation in cervical cancer by using diffusion-weighted MR imaging: a pilot study Ying Liu1, Haoran Sun2, Renju Bai2 and Zhaoxiang Ye1*
Abstract Background: To investigate the feasibility of DWI in evaluating early therapeutic response of uterine cervical cancer to concurrent chemoradiation (CCR) and establish optimal time window for early detection of treatment response. Methods: This was a prospective study and informed consent was obtained from all patients. Thirty-three patients with uterine cervical cancer who received CCR underwent conventional MRI and DWI examinations prior to therapy (base-line) and at 3 days (postT1), 7 days (postT2), 14 days (postT3), 1 month (postT4) and 2 months (postT5) after the therapy initiated. Tumor response was determined by comparing the base-line and postT5 MRI by using RECIST criterion. Results: Percentage ADC change (γADC) of complete response (CR) group at each follow up time was greater than that of partial response (PR) group, and the differences were significant at postT3 (p = 0.007), postT4 (p = 0.001), and postT5 (p = 0.019). There was positive correlation between γADC at each follow-up time and percentage size reduction at postT5. The day of 14 after the therapy initiated can be considered as the optimal time for monitoring early treatment response of uterine cervical cancer to CCR, and the representative and sensitive index was γADC. With the cut-off value of 35.4 %, the sensitivity and specificity for prediction of CR group were 100 % and 73.1 %, respectively. Conclusions: It is feasible to use DWI to predict and monitor early treatment response in patients with uterine cervical cancer that undergoing CCR, and optimal time window for early detection of tumor response is the day of 14 after therapy initiated.
Background Uterine cervical cancer remains to be the second most prevalent gynaecological tumors worldwide [1]. Recently, cisplatin-based concurrent chemoradiation (CCR) has been established as being more effective than radiation therapy (RT) alone because chemotherapy has been shown to increase the sensitivity of tumor cells to radiation and to control both local and systemic disease manifestations [2]. However, not all tumors of the same pathological type and FIGO stage will follow the same course of disease; individual responses to CCR have been * Correspondence:
[email protected] 1 Department of Radiology, Tianjin Medical University Cancer Institute and Hospital, National Clinical Research Center of Cancer, Key Laboratory of Cancer Prevention and Therapy, Tianjin, China Full list of author information is available at the end of the article
shown to vary significantly. Therefore, precise assessing tumor response before or at an early stage of CCR is important for treatment planning and for determining the prognosis. Tumor response to therapy is conventionally assessed by changes in tumor size according to the Response Evaluation Criteria in Solid Tumors (RECIST) using computed tomography (CT) or Magnetic resonance (MR) imaging. Unfortunately, the information from post-therapy imaging after treatment completion usually comes too late to realistically impact patient management [3]. If therapy failure can be predicted effectively and as early as possible in the initial course of treatment, it will provide a window of opportunity to change the initial therapy approach upfront and clinical management can be profoundly impacted in the individual patient [3]. Given this difficulty,
© 2015 Liu et al. Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.
Liu et al. Radiation Oncology (2015) 10:185
increasing demands are placed on imaging modalities to provide an early and reliable response marker which would have great clinical importance for uterine cervical cancer patients to cease ineffective treatment and to avoid delays in starting alternative, potentially more effective treatments. Diffusion-weighted MR imaging (DWI) is a functional MR imaging technique that can explore the random diffusion motion of water molecules in vivo [4]. Previous studies in a variety of tumor types have suggested that quantitative interpretation of apparent diffusion coefficient (ADC) can be used as a biomarker for response to treatment [5–9]. To date, a few clinical studies on the usefulness of DWI as a measurement of treatment response in uterine cervical cancer have been reported [10–13]. However, systematic investigation to evaluate the optimal time window to detect early response of tumor to CCR is lacking. The present study was therefore designed to systematically analyze dynamic changes of ADC after initiation of CCR, and determine whether ADC measurements of uterine cervical cancer before and after early initiation of CCR can be used to follow treatment response, in particular, to provide a timewindow of early detection of response to CCR.
Methods Patient population
Our study received institutional ethics committee approval (Tianjin Medical University General Hospital) and written informed consent was obtained from all patients. Forty-five patients with biopsy-proven uterine cervical cancer, who planned to receive CCR, were prospectively recruited to this study. Inclusion criteria consisted of [a] histologically (biopsy) proven squamous cell carcinoma of uterine cervix before the first MR examination, and the time interval between biopsy and baseline MR examination did not exceed 1 month; [b] FIGO stage based on clinical examination ranges from II to IV; [c] no previous radiation or CCR treatment for uterine cervical cancer; [d] treatment consisting of radiotherapy and cisplatin-based chemotherapy; [e] no contraindications for MR examination. Exclusion criteria consisted of [a] unable to complete the full course of treatment, [b] time interval between base-line MRI and start of treatment is more than a week, [c] fail to complete the followup MRI examinations on time. All participants were scheduled to receive six MR examinations: before CCR (base-line), at 3 days (postT1), 7 days (postT2), 14 days (postT3), 1 month (postT4) and 2 months (postT5) after therapy initiated. 12 patients were excluded from the study because of unable to complete the full course of treatment (5 patients) or fail to complete the follow-up MRI examinations on time owing to patient incompliance (7 patients). Finally, 33 female patients (mean age
Page 2 of 8
53.6 years; age range, 36–75 years) with uterine cervical cancer enrolled in this study. Treatment
All patients were scheduled to undergo external beam radiation therapy (EBRT) of the pelvis and intracavitary brachytherapy (ICBT). Treatment was composed of 2 days per week of brachytherapy in the form of intracavitary (60 Gy/12 fractions), and then 3 days per week of pelvic external beam radiotherapy (42 Gy/21 fractions to point B), accompanied with cisplatin chemotherapy at a dose of 40 mg/m2 during the intervening weekends. MR examination
All MR examinations were performed using a 1.5-T unit (Twin Excite, GE Healthcare, USA) with a torso phasedarray body coil. Before DWI, conventional T2-weighted fast spin-echo in the sagittal and transverse planes (TR/ TE, 4,000 ms/85 ms; matrix size, 320 × 224; band width, 31.25 Hz/pixel; field of view, 36 cm; number of excitations, 2; slice thickness, 6 mm; gap, 1 mm), T2-weighted fast spin-echo with fat suppression in the transverse plane (the parameters were the same as for the T2weighted image) and T1-weighted spin-echo in the transverse plane (TR/TE, 500 ms/20 ms; matrix size, 320 × 160; band width, 31.25 Hz/pixel; field of view, 36 cm; number of excitations, 2; slice thickness, 6 mm; gap, 1 mm) were obtained. Diffusion-weighted MR images were acquired using a non-breath-hold single-shot spin-echo echo-planar imaging (EPI) sequence and array spatial sensitivity encoding technique (ASSET) in the transverse plane (TR/TE, 4,000 ms/58.5 ms for b values of 0 and 1000 s/mm2; matrix size, 128 × 128; field of view, 36 cm; number of excitations, 4; slice thickness, 6 mm; gap, 1 mm; R factor, 2; phase-encoding direction, anteroposterior). The diffusionweighting gradients were applied in all three orthogonal directions. The scanning time of DWI was 1 min and 4 s. MR image analysis
MR images were analyzed by two radiologists who performed tumor ADC measurements and longest tumor diameter measurement on the pre- and post-CCR images independently. The readers were blinded to each other’s results. The longest tumor diameter was measured using the transverse plane on T2-weighted images. ADC maps were calculated on a pixel-by-pixel basis by using built-in software (AW4.3 Functool; GE Healthcare). For ADC calculation, up to three slices depicting the largest tumor diameter were selected and then tumor margins were free-hand delineated on DW images. In each slice a region of interest (ROI) was delineated according to the tumor geometry. The border of the ROI was placed in the tumor periphery close to the
Liu et al. Radiation Oncology (2015) 10:185
tumor margin in order to encompass as much of the tumor as possible with exclusion of hemorrhagic or necrotic areas. If the residual tumor was tiny and covered less than three slices, the ROIs were assessed twice in the same site by each experienced radiologist. In case of invisible residual tumors on MR images after completion of therapy at T5, the ROI was drawn on the same area where tumors were identified at base-line MR images. Mean ADC value was calculated by average for each tumor. The ADC changes (γADC) (as a percentage) for each lesion between the base-line and follow-up time points were calculated using the formula: γADC ¼ ðADCF −ADCB Þ=ADCB 100; where ADCB represents pretreatment ADC values, and ADCF represents the post-treatment ADC values after CCR initiated. The longest tumor diameter changes (γD) (as a percentage) for each lesion between the base-line and follow-up time points were calculated using the formula: γD ¼ ðDB −DF Þ=DB 100; where DB represents pretreatment longest tumor diameter, and DF represents the post-treatment longest tumor diameter after CCR initiated. Treatment outcome analysis
According to Response Evaluation Criteria in Solid Tumors (RECIST 1.1), response to treatment was determined by comparing the base-line MRI and follow-up MRI at 2 months after the therapy initiated [14]. Complete response (CR) was concluded if there was no residual tumor on T2-weighted images; partial response (PR) was concluded if the longest diameter of the tumor was less than 70 % of the original size; the disease was determined to be stable (SD) if there was neither sufficient shrinkage to qualify for partial response nor sufficient increase to qualify for progressive disease; and progressive disease (PD) was concluded if there was at least a 20 % increase in the longest diameter of tumor, taking as reference the longest diameter recorded pre-treatment. Statistical analysis
Statistical analyses were performed using the Statistical Package for the Social Sciences (SPSS, version 13.0). Intra-class correlation (ICC) of coefficient between two readers was calculated. Multiple comparisons of continuous data were performed by Randomized blocks analysis of variance. In order to test differences between two independent groups, statistical comparisons were made using Student’s t-test. Pearson correlation coefficient was used in order to test independence between
Page 3 of 8
variables. A ROC analysis was performed to determine the optimal time window for early detection of tumor response to CCR. From the ROC analysis, Youden’s index was used in determining the optimal threshold value. The resulting threshold values were then used to calculate the sensitivity and specificity. A level of p value
